Home

Search

Collections

Journals

About

Contact us

My IOPscience

Device optimization of tris-aluminum (Alq3) based bilayer organic light emitting diode structures

This article has been downloaded from IOPscience. Please scroll down to see the full text article. 2006 Smart Mater. Struct. 15 S92 (http://iopscience.iop.org/0964-1726/15/1/015) View the table of contents for this issue, or go to the journal homepage for more

Download details: IP Address: 142.51.1.212 The article was downloaded on 07/03/2013 at 06:24

Please note that terms and conditions apply.

INSTITUTE OF PHYSICS PUBLISHING

SMART MATERIALS AND STRUCTURES

Smart Mater. Struct. 15 (2006) S92–S98

doi:10.1088/0964-1726/15/1/015

Device optimization of tris-aluminum (Alq3) based bilayer organic light emitting diode structures J Chan1 , A D Raki´c1 , C Y Kwong2 , Z T Liu3 , A B Djuriˇsi´c3 , M L Majewski1 , W K Chan4 and P C Chui2 1

School of Information Technology and Electrical Engineering, The University of Queensland, Brisbane, QLD 4072, Australia 2 Department of Electrical and Electronic Engineering, University of Hong Kong, Pokfulam Road, Hong Kong 3 Department of Physics, University of Hong Kong, Pokfulam Road, Hong Kong 4 Department of Chemistry, University of Hong Kong, Pokfulam Road, Hong Kong E-mail: [email protected]

Received 10 May 2005 Published 13 December 2005 Online at stacks.iop.org/SMS/15/S92 Abstract In this work we present a detailed analysis of the emitted radiation spectrum from tris(8-hydroxyquinoline) aluminum (Alq3 ) based bilayer organic light emitting diodes (OLEDs) as a function of the choice of cathode, the thickness of the organic layers, and the position of the hole transport layer/Alq3 interface. The calculations fully take into account dispersion in the glass substrate, the indium tin oxide anode, and in the organic layers, as well as the dispersion in the metal cathode. The influence of the incoherent transparent substrate (1 mm glass substrate) is also fully accounted for. Four cathode structures have been considered: Mg/Ag, Ca/Ag, LiF/Al, and Ag. For the hole transport layer, N ,N
-diphenyl–N ,N
-(3-methylphenyl)–1,1
-biphenyl– 4,4
-diamine (TPD) and N ,N
-di(naphthalene-1-yl)–N ,N
-diphenylbenzidine (NPB) were considered. As expected, the emitted radiation is strongly dependent on the position of the emissive layer inside the cavity and its distance from the metal cathode. Although our optical model for an OLED does not explicitly include exciton quenching in the vicinity of the metal cathode, designs placing the emissive layer near the cathode are excluded to avoid unrealistic results. Guidelines for designing devices with optimum emission efficiency are presented. Finally, several different devices were fabricated and characterized and experimental and calculated emission spectra were compared. (Some figures in this article are in colour only in the electronic version)

1. Introduction Organic light emitting diodes (OLEDs) are attracting increasing attention for applications in display devices due to their high brightness and wide viewing angle [1, 2]. While there has been significant progress in this area over recent years, comprehensive optical studies on device simulation and design have been scarce. The simulations of the charge transport and charge distribution in organic light emitting 0964-1726/06/010092+07$30.00 © 2006 IOP Publishing Ltd

devices have been reported [3, 4]. There have been several works on calculating the emission pattern [5] and external coupling in OLEDs [6]. It has been shown that the emission from an OLED can be varied through interference effects by changing the thickness of the organic layer in an OLED with a dielectric layer located below the indium tin oxide (ITO) anode [7] or by changing the thickness of a ZnSe layer on top of the metal cathode [8, 9]. For an OLED with SiO2 layer below the ITO, it was shown that the emission spectrum

Printed in the UK

S92

Device optimization of tris-aluminum (Alq3 ) based bilayer OLED structures

2. Description of the model

2.45 3.1 Li F

Ag Alq3 HTL ITO glass glass

HTL

Alq3

Ag 4.3

ITO 4.6-5.0 5.3-5.6 5.8

Figure 1. The typical OLED device structure. Energy levels (in eV) of the materials are also shown.

from an OLED is strongly dependent on the Alq3 thickness, so that with the increase of Alq3 thickness the emission peak shifts from ∼525 to ∼555 nm, and with further increase of the thickness even splits into two peaks, one located ∼495 nm and the other ∼590–600 nm [7]. It was also shown that, by changing the thickness of the ZnSe layer deposited on top of the cathode, the efficiency of the OLED can be optimized so that for optimal ZnSe thickness efficiency as high as 64 cd A−1 can be achieved [8, 9]. The spectral and angular distribution of the emission from an OLED has been studied in more detail in organic microcavity devices [10, 11] than in non-cavity devices. However, if we consider a typical OLED structure consisting of a glass substrate, ITO anode, hole transport layer (HTL), emitting/electron transport layer (EML/ETL), and a metal cathode, as shown in figure 1, it is obvious that this device can be considered as a microcavity with a very weakly reflecting bottom mirror. It has been shown that weak microcavity effects can be observed in such simple bilayer OLEDs [12]. If we consider the refractive indices of the organic materials used [13], the ITO [14] anode, and the glass substrate, the reflectance from the bottom side is just a few per cent, but this is still sufficient to affect the emission pattern of an OLED and result in a dependence of the emission wavelength and intensity on the organic layer thickness and possibly the cathode choice. Therefore, in this work we investigated the influence of the cathode choice, organic layer thickness, and the position of the emission region to the output from an OLED consisting of N ,N
-diphenyl–N ,N
-(3-methylphenyl)– 1,1
-biphenyl–4,4
-diamine (TPD) or N ,N
-di(naphthalene-1yl)–N ,N
-diphenylbenzidine (NPB) as a hole transport layer (HTL) and tris(8-hydroxyquinoline) aluminum (Alq3 ) as emitting/electron transport layer (EML/ETL). The cathodes considered are the structures commonly used in OLEDs: Mg/Ag, Ca/Ag, LiF/Al, and Ag. The model fully takes into account the dispersion in all the layers in the structure. The organic layer thickness values chosen are between 30 and 250 nm, and the position of the emission region close to the metal cathode has not been considered in order to avoid unrealistic predictions since the model does not take into account exciton quenching due to defect states introduced during metal deposition. The calculations have been compared with the experimental data for ITO/NPB/Alq3 /LiF/Al and ITO/NPB/Alq3 /Ag devices. The paper is organized as follows. In the following section, the description of the model is given. In section 3, experimental conditions are described. The results and discussion are presented in section 4. Finally, conclusions are drawn.

This paper investigates the effect of cathode metal and the position of the emission region on the spectral and luminous characteristics of the OLED. We analyzed conventional OLEDs fabricated on ITO glass substrates as microcavity devices with ITO/glass considered as a bottom mirror. Due to the small index offset between the ITO and the adjacent HTL the cavity effects are weak. Resonant modes of the microcavity have to satisfy the condition that the phase change during one round trip is a multiple of 2π. In other words, for normal incidence, the following equation is valid:
4π i

λ

di n i (λ) − ϕtop (0, λ) − ϕbot (0, λ) = 2mπ,

(1)

where λ is the emission wavelength, ϕtop (λ), ϕbot (λ) are the wavelength dependent phase changes upon reflection from top and bottom mirrors, respectively, m is an integer which defines the mode number, and the summation is performed over all the layers inside the cavity with thicknesses di and refractive indices n i (λ). The phase shift upon reflection from the mirrors is calculated using the matrix method. The wavelength dependence of the index of refraction for aluminum (Al) and silver (Ag) was modeled using Brendel– Borman’s model, with model parameters given in [15]. The refractive index of magnesium was modeled by an oscillator (Lorentz–Drude) model with parameters given in [16], while the optical constants of calcium films were obtained from the recent work of Ramsdale and Greenham [17]. The refractive index of Alq3 and TPD was modeled using the Cauchy equation n(λ) = A + B/λ2 + C/λ4 , where the coefficients A, B, and C were determined by fitting the data in [13]. The refractive index of NPB was determined from the spectroscopic ellipsometry measurements of an NPB film on a Si substrate. Ellipsometric measurements were taken at the incident angles of 65◦ , 70◦ , and 75◦ , in the wavelength range 300–1000 nm. The index of refraction of NPB was modeled using a two-oscillator Lorentz model, and three incident angles were fitted simultaneously. The refractive index of ITO was modeled using the model given in [14]. As all devices investigated here are bottom emitting, the reflection from the back face of the glass substrate needs to be considered as well. The reflection coefficient was determined by applying matrix formalism with incoherent substrate correction using the modified matrix approach of Katsidis and Siapkas [18, 19]. The refractive index of glass was described using a Cauchy equation. The external emission spectrum of the OLED is given by [10, 11]:    1−Rbot  4π z i i 1 + Rtop + 2 Rtop cos( λ − ϕtop ) i  Icav (λ) = 1 + Rbot Rtop − 2 Rbot Rtop cos( 4πλL − ϕtop − ϕbot ) × Inc (λ), (2)  where L = j n j d j is the optical thickness of the cavity, z i is the optical distance of the emitting dipoles to the metal mirror (cathode) with reflectivity Rtop , Rbot is the reflectivity of bottom mirror (ITO/glass), Inc (λ) is the free space emission of the emitting material, and the summation over i is performed assuming that the dipoles are located within 10–20 nm from the HTL/emitting layer interface [10]. The actual thickness of the emitting layer depends on the exciton diffusion length, S93

J Chan et al 3000

140

Mg/Ag

Cavity thickness 80 nm 100 nm 120 nm 140 nm 160 nm

Luminance Spectrum (A./U.)

2500 120 2000 100 1500 80 1000

500

0 400

500

600

Ag

120

2000

100

1500

80

500

0 400

700

500

700

3000 Cavity thickness 80 nm 100 nm 120 nm 135.2 nm 160 nm

Ca/Ag

LiF/Al

Cavity Thickness 80nm 100nm 120nm 160nm 180nm 151nm

120

2500

2000 100

1500 80

1000

500

500

600

700

Wavelength (nm)

2500

Luminance Spectrum (A./U.)

Luminance Spectrum (A./U.)

600

Wavelength (nm)

3500

0 400

Cavity thickness 80 nm 100 nm 120 nm 140 nm 160 nm

1000

Wavelength (nm)

3000

140

2500

Luminance Spectrum (A./U.)

3000

120 100

2000

1500 80

1000

500

0 400

500

600

700

Wavelength (nm)

Figure 2. Modeled effect of the cavity length on the electroluminescent spectra of devices with different top metallic mirrors.

and therefore can be affected by the quality of organic layers (deposition rate, purity of materials, vacuum level, interface surface roughness), as well as the driving current and environmental factors since organic materials are very sensitive to temperature and atmosphere exposure.

3. Experimental details The materials used (Alq3 from Syntech GmbH, and NPB from H W Sands Corp.) were purified by vacuum sublimation before fabrication. ITO substrates with nominal sheet resistance of 40 /square were supplied by Varitronix Limited, Hong Kong. All the ITO substrates were cleaned before fabrication with the following cleaning procedure. The substrates were first rinsed by ethanol, acetone, toluene, then acetone, ethanol, and deionized (DI) water. After that, the substrates were cleaned in ultrasonic bath for 10 min in acetone, ethanol, and DI water, and dried in an oven at around 90 ◦ C. We have performed optimal surface treatment of the cleaned ITO according to previously reported results [20] by dipping the ITO into 12% aqueous solution of HCl for 15 min, followed by 15 min UV– ozone treatment. Devices were fabricated by vacuum deposition. The pressure during evaporation was of the order 10−4 Pa. The evaporation rate was kept at 1–2 Å s−1 . The distance from S94

source to film was about 23 cm to ensure uniformity of film thickness, and the substrate holder was rotating. The thickness of the films was controlled by a quartz thickness monitor. After evaporation, the film thickness was verified using a Dektak 3 step profiler. The electrode consisted of 0.5 nm LiF and 100 nm Al. Eight cells were fabricated on each ITO substrate in one fabrication, and fabrications were repeated several times to verify the reproducibility of the obtained results. The electroluminescence (EL) was determined using luminance– current–voltage (L–I –V ) characteristics that were measured in ambient environment using a Keithley 2400 sourcemeter and Oriel MS 260I monochromator with integrating sphere and CCD detector. Photoluminescence (PL) spectra were recorded using a Perkin Elmer Instruments LS50B Luminescence Spectrometer for an excitation wavelength of 350 nm.

4. Results and discussion As we have mentioned before, the cavity effects in structures used in this study (shown in figure 1) are relatively weak. The substrate used is 1 mm thick glass whereas the anode is a 50 nm thick ITO layer. We investigated four different metallic cathode structures, namely: LiF 0.5 nm/Al 100 nm, Mg 30 nm/Ag 70 nm, Ag 80 nm, and Ca 30 nm/Ag 70 nm. Figure 2 shows the calculated effect of the cavity length on the

Device optimization of tris-aluminum (Alq3 ) based bilayer OLED structures

EL Intensity (A./U.)

(a)

LiF/Al Mg/Ag Ag Ca/Ag

2800

2400

Ca/Ag 2000

Ag 1600

Mg/Ag

LiF/Al

80

100

120

140

160

Cavity Thickness (nm) 3000

LiF/Al Ag Mg/Ag Ca/Ag

2500

EL Intensity (a./u.)

electroluminescent spectra of devices with these four different top metallic mirrors. All four structures show similar patterns of dependence on the cavity thickness: monotonic increase in extracted EL intensity with increasing cavity thickness (solid lines in figures 2(a)–(d)) up to the value corresponding to the first longitudinal mode of the cavity. Beyond the resonant thickness the intensity is decreasing (lines with markers). Structures with LiF/Al, Mg/Ag and Ag mirrors have similar optical behavior due to the strong free-electron-like behavior of these metals. The device with Ca/Ag mirror has significantly reduced EL intensity due to the lower reflectivity of calcium in comparison to the other cathodes used. In all cases strong enhancement in extracted EL intensity was obtained for cavity lengths corresponding to the resonance of the first longitudinal mode of the cavity given by equation (1), notwithstanding the fact that the bottom mirror has very low reflectivity. In all calculations we assumed physical thicknesses of the HTL (TPD) and the ETL (Alq3 ) layers equal to 0.5 of the cavity length. Figure 3(a) shows the calculated peak EL intensity as a function of the cavity thickness. The first three devices have the optimal thickness close to 140 nm, while the Ca/Ag cathode structure allows for a reduced cavity thickness. Figure 3(b) shows the magnitude of the EL peak as a function of Alq3 layer thickness. It should be noted that our calculations deal with optical factors only. The actual EL intensity will also depend on the charge transport properties in the organic layers and the electron injection efficiency of the cathode. In this work, we mainly study the optical effects of changing the layer thickness, interface position, and the cathode material. In future work, the optical design considerations will be combined with the electronic design issues to present comprehensive guidelines for efficient OLED fabrication. Figure 4 shows the dependence of the extracted EL spectrum on the position of the light emitting layer. For all four structures, we used the total cavity length corresponding to the first longitudinal mode (m = 1); we changed the thickness of the ETL (Alq3 ) to investigate the effect of the position of the light emitting layer on the EL spectrum. Thin, electrically unsuitable layers (thinner than 30 nm) have been avoided in calculations. The difference between the refractive indices of TPD and Alq3 is less than 1% in the vicinity of 530 nm (position of the EL peak of the non-cavity device). Therefore, the total physical cavity length is almost independent of the position of the TPD/Alq3 interface. This enabled us to single out and investigate the effect of the position of the emitting layer on the EL spectrum, without altering the other parameters of the optical design of the cavity. However, one has to bear in mind that changing the relative thicknesses of HTL and ETL (TPD and Alq3 respectively) electron transport will be significantly affected due to the different carrier mobility in the two layers. The mobility of electrons in Alq3 is about two orders of magnitude lower than the hole mobility in TPD [21]. Therefore, changing the thickness of these two layers will significantly affect the balance between electrons and holes reaching the interface, and thus affect the emission intensity. We will now have a closer look at the LiF/Al mirror device, since LiF/Al is the most commonly used cathode in OLEDs. The thickness of the Alq3 layer was changed from 30 to 110 nm. The total thickness of the cavity spacer was

(b)

Ag

2000

Ca/Ag 1500

Mg/Ag

1000

LiF/Al 500 20

40

60

80

100

120

Alq3 layer thickness (nm)

Figure 3. (a) Modeling of peak EL intensity as a function of the cavity thickness for different metal cathodes. (b) Modeling of peak EL intensity as a function of the Alq3 layer thickness for different metal cathodes.

kept at 151 nm, corresponding to the wavelength of the m = 1 resonant mode of the cavity. The effect of the position of the TPD/Alq3 interface on the electroluminescence spectrum in an OLED with LiF/Al mirror is shown in figure 4(a). Clearly the strongest EL enhancement is obtained when the emission layer is aligned with the position of the antinode of the resonant cavity [22]. Figure 5 depicts the electric field intensity in the optimally designed cavity with the HTL/ETL interface (emission layer) located at the antinode of the electric field. The phase shift on reflection from the metal mirror plays the critical role in determining the position of the antinode of the electric field. It is well known that the optical constants of metallic films depend strongly on the evaporating conditions and evaporation rate [15]. The optical constants of aluminum in particular may vary significantly due to the variable surface roughness of the film and due to the rapid volume oxidation of the aluminum film, unless the evaporation was performed in ultrahigh vacuum. It is usually assumed that the light emission occurs from the 20 nm thick layer of Alq3 adjacent to the TPD/Alq3 interface [10]. Figure 6 shows the calculated dependence of the EL intensity and EL peak wavelength as a function of the emission region thickness. To investigate the effect of the thickness of this region we selected two devices with interface S95

J Chan et al 3000

3000 Interface position 30/121 40/111 50/101 60/91 70/81 80/71 90/61 100/51 110/41

Luminance Spectrum (A./U.)

2500

2000 50/101

1500 40/111

1000

500 30/121

0 400

500

600

2500

2000

1500 30/110

1000

500

0 400

700

500

600

700

Wavelength (nm)

Wavelength (nm) 3000

3000

2500

2000

1500

1000 30/112

500

0 400

500

600

Interface position 30/105.2 nm 40/95.2 nm 50/85.2 nm 60/75.2 nm 70/65.2 nm 80/55.2 nm 90/45.2 nm 100/35.2 nm

Ca/Ag 2500

Luminance Spectrum (A./U.)

Interface position 30/112 40/102 50/92 60/82 70/72 80/62 90/52 100/42 110/32

Ag Luminance Spectrum (A./U.)

Interface position 30/110 40/100 50/90 60/80 70/70 80/60 90/50 100/40 110/30

Mg/Ag

Luminance Spectrum (A./U.)

LiF/Al

70/65 2000

60/75

1500

50/85

30/105

500

0 400

700

40/95

1000

500

Wavelength (nm)

600

700

Wavelength (nm)

Figure 4. Calculated dependence of the extracted EL spectrum on the position of the light emitting layer for different metal cathodes.

1.0

3500

ITO

Alq3

TPD

LiF Al

75 Alq3

3000

Luminance Spectrum (A./U.)

Electric Field Intensity (A.U)

0.8

0.6

0.4

0.2

2500

76 nm TPD / 75nm Alq3 Emission layer thickness 20nm 10nm 111 TPD 40 Alq3 20nm 10nm

5 nm 5nm

2000

1500

40 Alq3

1000

500

0.0

50

148

201

301.5

OLED layers (nm)

Figure 5. Electric field intensity of optimally designed cavity.

positions (emitting region positions) located on the opposite sides of the standing wave maximum. The thickness of the emission region within the Alq3 film was varied between 5 and 20 nm. By reducing the thickness of the emission region of the 111 nm TPD/40 nm Alq3 structure the position of the emission region is effectively better aligned with the antinode S96

0 400

500

600

700

800

Wavelength (nm)

Figure 6. Dependence of the EL intensity and EL peak wavelength as a function of the emission layer thickness.

position of the field. The other structure (76 nm TPD/75 nm Alq3 ) is effectively misaligned by the reduction on the emission region thickness. Therefore, to maximize the emitted power, the antinode of the optical field should be aligned with the middle of the emission region. This makes the thickness of

Device optimization of tris-aluminum (Alq3 ) based bilayer OLED structures 2.4

Electroluminescence Spectrum (a./u.)

2.2 2.0 1.8 1.6

n, NPB k, NPB n, Alq

1.4 1.2 1.0

3

0.8

k, Alq3

0.6

n T PD, Cauchy model n Alq, Cauchy model

0.4 0.2 0.0 400

500

600

700

Wavelength (nm)

Model 35.4nm NPB/250nm Alq3 Model 70.4nm NPB/166.8 nm Alq3 Exp 35.4nm NPB/250nm Alq3 Exp 70.4nm NPB/166.8 nm Alq3

0.15

0.10

0.05

0.00 400

500

600

700

800

Wavelength (nm)

Figure 7. The real and imaginary parts of the refractive index of Alq3 and NPB. The refractive index from the Cauchy model for Alq3 and TPD is shown for comparison.

Figure 8. Comparison between experimental and calculated electroluminescence spectra for different thicknesses of the organic layers.

the Alq3 layer in the optimally designed device dependent on the thickness of the light emission region. As pointed out before, the emission region thickness is determined by the exciton diffusion length. The exciton diffusion length can be affected by the experimental parameters, such as the purity of material used, vacuum level during deposition, deposition rate, etc, as well as experimental conditions during measurement, such as bias, temperature, and measurement environment (with or without encapsulation, in atmosphere or in inert gas). Since some of the experimental factors may be difficult to control in a reliable and reproducible manner, it is difficult to ensure that all the fabricated devices would have the same emission region width. However, it is possible to artificially confine the excitons by insertion of an additional wide band gap organic layer such as bathocuproine, which is typically used in organic light emitting diodes and organic solar cells as a hole blocking or exciton blocking layer [23, 24]. In such a manner, the width of the emission region can be made sufficiently narrow and very reproducible. The width of the emission region can also be reduced by doping (with low dopant concentrations) [25]. Figure 7 shows the refractive indices of the organic materials considered (Alq3 as emission layer and TPD or NPB as hole transport layer). The small difference between our result and the Cauchy model for the refractive index of Alq in [13] is likely due to different source and purity of the organic materials, and different procedures for extraction of the experimental data. While these materials have similar refractive indices in the spectral region of interest, and thus their optical properties will not affect the device performance, there are other factors which need to be considered in order to fabricate efficient OLED devices. As already mentioned, there are significant differences in the carrier mobility [21]. In addition, the choice of the hole transport layer affects the hole injection efficiency from the ITO electrode into the organic material [26]. Also, the higher glass transition temperature of NPB (90–100 ◦ C, compared to 60–70 ◦ C for TPD [27]) results in improved stability. Higher luminous efficiency for the same device structure can be obtained when NPB is used instead of a TPD layer [20]. Thus, although calculation results are valid for devices using both conventional hole transport layers

due to their similar optical properties, for the comparison with the experimental data, devices with NPB as an HTL were fabricated to ensure high brightness and sufficient lifetime for detailed characterizations such as the angular dependence of the emission. Figure 8 compares the calculated and experimental EL spectra for two devices fabricated with a LiF/Al mirror and with two different thicknesses of the Alq3 layer: 250 and 167 nm. The spectra reveal shifts in peak positions similar to those predicted by the optical model. We have chosen to model the EL spectra since in the case of PL the emission region is not as clearly determined as in the case of EL. For EL, excitons are created at the NPB/Alq interface and the emission region width is determined by the exciton diffusion length. For PL, the excitons can be generated throughout the Alq layer, but we cannot analytically account for the quenching near the interface with the metal cathode. Thus, a comparison between EL and the calculated spectra is more realistic. Excellent agreement can be observed between the calculated and the experimental peak positions. The experimental spectra, however, exhibit larger broadening than the calculated spectra. This is likely due to some complex experimental factors, such as enhanced Joule heating around pinhole or defect areas, emissive layer degradation due to formation of unstable Alq3 cation species, or reaction with oxygen and water in the atmosphere, etc, which may affect the obtained results. In order to obtain better agreement with the experimental data, it would be necessary to fully take into account the charge injection, charge transport, and exciton diffusion in modeling the emitted spectra. Work is in progress to develop a comprehensive model taking into account both electronic and optical properties of all the layers in the device.

5. Conclusion In conclusion, we have presented calculations of the emitted radiation from organic OLEDs as a function of the choice of cathode, the thickness of the organic layers, and the position of the hole transport layer/Alq3 interface. Four cathode structures have been considered: Mg/Ag, Ca/Ag, LiF/Al, and Ag. For the S97

J Chan et al

hole transport layer, N ,N
-diphenyl–N ,N
-(3-methylphenyl)– 1,1
-biphenyl–4,4
-diamine (TPD) and N ,N
-di(naphthalene1-yl)–N ,N
-diphenylbenzidine (NPB) were used, while tris(8hydroxyquinoline) aluminum (Alq3 ) was the emitting layer. It was found that the emission spectrum is strongly dependent on the layer thickness and interface position, as well as the thickness of the emission region. For the optimal efficiency, the emission region should be narrow and its center aligned with the antinode of the electric field in the resonant cavity. Since the emission region width depends on the exciton diffusion length, in order to confine the excitons to a narrow region the insertion of an additional exciton blocking layer or doping of the emitting layer may be advisable. Our simulation results show good agreement with the experiment. The objective of our work was to analyze an OLED structure from an optical point of view, and electronic considerations (electron injection and charge transport) were not taken into account. In order to obtain better agreement with electroluminescence experimental data, it will be necessary to integrate the electrical and optical models. Inclusion of electron injection efficiency for different cathodes and the charge transport in the organic layers will also ensure better prediction of electroluminescence intensity for devices with different cathodes.

Acknowledgments This work has been partially supported by the RGC earmarked grant of Hong Kong (Project HKU 7056/02E) and the University of Hong Kong University Research Committee seed funding grant.

References [1] Br¨utting W, Berleb S and Muckl A G 2001 Device physics of organic light-emitting diodes based on molecular materials Org. Electron. 2 1–36 [2] Hung L S and Chen C H 2002 Recent progress of molecular organic electroluminescent materials and devices Mater. Sci. Eng. R 39 143–202 [3] Crone B K, Davids P S, Campbell I H and Smith D L 2000 Device model investigation of bilayer organic light emitting diodes J. Appl. Phys. 87 1974–82 [4] Tutiˇs E, Berner D and Zuppiroli L 2003 Internal electric field and charge distribution in multilayer organic light-emitting diodes J. Appl. Phys. 93 4594–602 [5] Kim J-S, Ho P K H, Greenham N C and Friend R H 2000 Electroluminescence emission pattern of organic light-emitting diodes: Implications for device efficiency calculations J. Appl. Phys. 88 1073–81 [6] Lu M-H and Sturm J C 2002 Optimization of external coupling and light emission in organic light-emitting devices: modeling and experiment J. Appl. Phys. 91 595–604 [7] So S K, Choi W K, Leung L M and Neyts K 1999 Interference effects in bilayer organic light-emitting diodes Appl. Phys. Lett. 74 1939–41 [8] Riel H, Karg S, Beierlein T, Rieß W and Neyts K 2003 Tuning the emission characteristics of top-emitting organic light-emitting devices by means of a dielectric capping layer: An experimental and theoretical study J. Appl. Phys. 94 5290–6

S98

[9] Riel H, Karg S, Beierlein T, Ruhstaller B and Rieß W 2003 Phosphorescent top-emitting organic light-emitting devices with improved light outcoupling Appl. Phys. Lett. 82 466–8 [10] Dodabalapur A, Rothberg L J, Jordan R H, Miller T M, Slusher R E and Phillips J M 1996 Physics and applications of organic microcavity light emitting diodes J. Appl. Phys. 80 6954–64 [11] Shiga T, Fujikawa H and Taga Y 2003 Design of multiwavelength resonant cavities for white organic light emitting diodes J. Appl. Phys. 93 19–22 [12] Bulovi´c V, Khalfin V B, Gu G, Burrows P, Garbuzov D Z and Forrest S R 1998 Weak microcavity effects in organic light-emitting devices Phys. Rev. B 58 3730–40 [13] Celii F G, Harton T B and Philips D F 1997 Characterization of organic thin films for OLEDs using spectroscopic ellipsometry J. Electron. Mater. 26 366–71 [14] Laux S, Kaiser N, Z¨oller A, G¨otzelmann R, Lauth H and Bernitzki H 1998 Room-temperature deposition of indium tin oxide thin films with plasma ion-assisted evaporation Thin Solid Films 335 1–5 [15] Raki´c A D, Djuriˇsi´c A B, Elazar J M and Majewski M L 1998 Optical properties of metallic films for vertical-cavity optoelectronic devices Appl. Opt. 37 5271–83 [16] Machorro R, Siqueiros J M and Wang S 1995 Optical properties of Mg, from UV to IR, using ellipsometry and reflectometry Thin Solid Films 269 1–5 [17] Ramsdale C M and Greenham N C 2003 The optical constants of emitter and electrode materials in polymer light-emitting diodes J. Phys. D: Appl. Phys. 36 L29–34 [18] Katsidis C C and Siapkas D I 2002 General transfer-matrix method for optical multilayer systems with coherent, partially coherent, and incoherent interference Appl. Opt. 41 3978–87 [19] Mitsas C L and Siapkas D I 1995 Generalized matrix-method for analysis of coherent and incoherent reflectance and transmittance of multilayer structures with rough surfaces, interfaces, and finite substrates Appl. Opt. 34 1678–83 [20] Li C N, Djuriˇsi´c A B, Kwong C Y, Lai P T, Chan W K and Liu S Y 2005 Indium tin oxide surface treatments for improvement of organic light emitting diode performance Appl. Phys. A 80 301–7 [21] Naka S, Okada H, Onnagawa H, Yamaguchi Y and Tsutsui T 2000 Carrier transport properties of organic materials for EL device operation Synth. Met. 111–112 331–3 [22] Yamamoto Y, Machida S, Igeta K and Bjork G 1991 Controlled spontaneous emission in microcavity semiconductor lasers Coherence, Amplification and Quantum Effects in Semiconductor Lasers ed Y Yamamoto (New York: Wiley) [23] Li F, Zhang M, Feng J, Cheng G, Wu Z, Ma Y, Liu S, Sheng J and Lee S T 2003 Red electrophosphorescence devices based on rhenium complexes Appl. Phys. Lett. 83 365–7 [24] Peumans P and Forrest S R 2001 Very-high-efficiency double-heterostructure copper phthalocyanine/C60 photovoltaic cells Appl. Phys. Lett. 79 126–8 [25] Kalinowski J, Picciolo L C, Murata H and Kafafi Z H 2001 Effect of emitter disorder on the recombination zone and the quantum yield of organic electroluminescent devices J. Appl. Phys. 89 1866–74 [26] Giebeler C, Antoniadis H, Bradley D D C and Shirota Y 1999 Influence of the hole transport layer on the performance of organic light-emitting diodes J. Appl. Phys. 85 608–15 [27] Carrard M, Conto-Gonclaves S, Si-Ahmed L, Ades D and Siove A 1999 Improved stability of interfaces in organic light emitting diodes with high Tg materials and self-assembled monolayers’ Thin Solid Films 352 189–94